CC BY-NC-ND 4.0 · Journal of Neuroanaesthesiology and Critical Care 2017; 04(04): S49-S55
DOI: 10.4103/jnacc-jnacc-75.16
Conference Proceeding
Thieme Medical and Scientific Publishers Private Ltd.

Critical care of subarachnoid haemorrhage

Michael J. Souter
1  Department of Anesthesiology and Pain Medicine, University of Washington School of Medicine, Harborview Medical Center, Seattle, WA, USA
› Author Affiliations
Further Information

Address for correspondence:

Prof. Michael J. Souter
Department of Anesthesiology and Pain Medicine
University of Washington School of Medicine, Harborview Medical Center, 325 Ninth Avenue, Seattle, WA

Publication History

Publication Date:
08 May 2018 (online)



Subarachnoid haemorrhage (SAH) is a consistent presentation of haemorrhagic stroke of significance to clinicians in neurocritical care, inducing consequent effects on non-neurological systems, while at the same time, rendering the brain vulnerable to secondary physiological insult modifying neurological outcome, despite control of the original point of haemorrhage. Coordinated treatment depends on comprehensive evaluation of both cerebral and systemic physiology, identifying and treating impaired function. The presence of a dedicated neurocritical care team can benefit outcome. Protocols of care have evolved to meet evidence-based challenges, discarding potentially deleterious components of hypervolaemia and haemodilution, while maintaining pressure-guided perfusion. Treatment targets have also evolved with a shift in focus away from SAH-associated vasospasm, towards actual ischaemic outcome – illustrated by lack of effectiveness of pharmaceutical treatments of vasospasm. Clinicians must consequently review pathophysiological mechanisms of injury and devise new treatment opportunities.



Aneurysmal subarachnoid haemorrhage (aSAH) is a subtype of haemorrhagic stroke of sustained relevance to neuroanaesthesia and neurocritical care across countries and continents, accounting for 3% of all strokes.[1] The incidence of SAH ranges from 2 to 22 patients per 100,000 population per year, with greatest numbers in Finland and Japan.[2] [3] [4] Mortality ranges from 26% to 50%, with a 21% risk of death in first 24 h after haemorrhage, increasing to 37% and 44% at 7 and 30 days, respectively.[5] The pre-hospital mortality of 12%–15% cannot be ignored and may contribute to greater hospital mortality as detection, primary support and referral mechanisms evolve.[2]

Risk factors for aneurysmal haemorrhage include a family history of aSAH (especially in a first-degree relative), hypertension, tobacco smoking, use of cocaine (and other sympathomimetics), having a known aneurysm >7 mm in size and presence of an aneurysm in the posterior communicating artery or in the posterior circulation.[2]



There is a 1%–6% incidence of previously unidentified aneurysms detected at autopsy and some familial clustering of aSAH. There are also identified connective tissue disorders well associated with higher risk, for example, Ehlers–Danlos syndrome, neurofibromatosis type I, Marfan syndrome and autosomal dominant polycystic kidney disease.[6] This led to some belief that aneurysms arise due to congenital defects in the arterial wall structure. Fine anatomic studies and improved insights into the degenerative consequences of atherosclerotic disease, flow dynamics and wall shear stress on the muscular layers of the artery have largely disproved this assumption, validated by the presence of aneurysms around points of branch turbulence and/or highly pulsatile oscillations of flow.[7]

Those interactions of flow and a pathologically weakened arterial musculature facilitate the formation and expansion of aneurysms to the point that they exceed vessel wall strength, with consequent rupture and haemorrhage into parenchyma or the subarachnoid space. Other causes of SAH are listed in [Table 1].

Table 1

Other causes of subarachnoid haemorrhage

Head trauma

Perimesencephalic haemorrhage

Arteriovenous malformation rupture

Dural arteriovenous fistula

Mycotic aneurysm

Cocaine, methamphetamine or other sympathomimetic abuse

Hypertensive crisis

Moyamoya syndrome

Central nervous system vasculitis

Call-Fleming syndrome (reversible cerebral vasoconstriction syndrome)

Intradural vertebral artery dissection

Pituitary apoplexy



Typically, patients with aSAH present with sudden onset of a severe headache, often described as the ‘worst headache of life.’ However, it should be noted that only 6%–17% of patients presenting with a sudden severe headache will, in fact, turn out to have aSAH.[8] [9] Other typical symptoms include seizure, loss of consciousness and sudden vomiting along with onset of headache [Table 2] [8]

Table 2

Common presenting features

Sentinel leaks - 30%-50%

Meningitis - 80%

Loss of consciousness - 45% at ictus, 10% for several days

Seizures - 10%-25%

Focal neurology - 25% (hemiparesis, aphasia, hemineglect, cranial nerve palsies and memory loss)

Motor deficits - 10%-15%

Retinal haemorrhage (Terson's syndrome) - 20%-30%

Hypertension - 50%

Up to 50% of patients describe a ‘sentinel headache’ or ‘herald bleed’ a similar sudden onset severe headache that presents around 3–7 days prior, but spontaneously resolves. This is thought due to small tears or ‘micro-leaks’ from aneurysm, which eventually rupture to cause the main haemorrhage.

Physical findings are often non-specific but may include depressed level of consciousness or confusion. A new third cranial nerve palsy may be associated with an ipsilateral posterior communicating artery aneurysm compressing or stretching the adjacent nerve.



Following the clinical presentation, diagnosis is usually confirmed by non-contrast head computed tomography (CT), which is highly sensitive to acute bleeding, but that sensitivity attenuates with time.[10]

If that initial head CT does not reveal SAH, a lumbar puncture should be performed. Xanthochromia will appear 6 h after haemorrhage or other cause of presentation and is best detected by spectrophotometry.[11]

Magnetic resonance imaging may also be employed where the initial head CT was negative, but there is a strong clinical history of >72 h duration.[12]

The gold standard remains digital angiography although CT angiography (CTA) has become increasingly helpful and can often be used more easily and quickly than catheter studies. Three-dimensional CTA allows simulated perspective and planning of the surgical approach although vessels <1 mm are not usually seen.[13]

The conventional cerebral angiogram, however, still provides the best definition of aneurysmal and vascular anatomy to facilitate surgical planning. If a first angiogram is negative, there is a 6% detection rate of an aneurysm on repeat angiography around 1 week later.[14]



Grading of the severity of aneurysmal haemorrhage allows comparison of severity, efficacy of care and outcomes. The Hunt and Hess (H and H) scale[15] and World Federation of Neurological Surgeons (WFNS) scale[16] are common clinical scales in use, while the Fisher grading scale is a radiologic scale that was used to predict the risk of vasospasm.[17] The H and H scale has been criticised for high degrees of inter-observer variability and relatively coarse discrimination of outcome.[16] The WFNS scale combines the Glasgow coma scale score with the presence or absence of focal deficits, with a more graduated relationship to outcome. The Prognosis on Admission of Aneurysmal Subarachnoid Haemorrhage scale is a relatively new score with possibly better prognostic discrimination.[18] The Fisher scale has been criticised for confusing outcome with progression through the stages of severity, where a grade III may have a worse outcome than a grade IV.



Treatment goals in aSAH are prioritised on, first, prevention of re-bleeding, followed by aneurysmal obliteration, with subsequent monitoring for delayed cerebral ischaemia (DCI).

There is a 4%–17% re-bleeding rate with aSAH in the first 72 h after presentation, with a mortality between 50% and 60%.[19] [20] [21] Anti-fibrinolytics were used historically but increased the risk of DCI[22] and hydrocephalus.[23] There is renewed interest in brief courses of <72 h where treatment of aneurysm is delayed by patient instability, but there are still concerns of hydrocephalus.[24]

Control of blood pressure has also been thought to reduce re-bleeding rates but with very little concrete evidence to support definitive blood pressure targets.[25] Suggested systolic targets range between 90 and 140 mmHg, while there is concern for increased risk of ischaemic stroke with aggressive reductions.[26]

Early treatment of aneurysm logically reduces concerns for blood pressure management and is performed by surgical clipping or endovascular coiling of aneurysm. Choice of treatment differs widely across the world with increasing endovascular use in Europe, but less so in the United States. Many factors may influence that decision including patient’s age, aneurysm location, shape and size, presence of haematoma, along with patient’s comorbidities and clinical state.



Acute hydrocephalus occurs in up to 20% of patients with aSAH,[27] with age, female gender, poor H and H grade on admission posterior location and Fisher grade 4, associated with increased risk.[28] Persistent headache, loss of upward gaze and cognitive impairment are typical, with that latter aspect often being confused with vasospasm. Those symptoms can be improved by drainage,[29] and any severely obtunded patient should be considered for a trial of external cerebrospinal fluid drainage before definitive prognostication or limitation of care.



DCI is the main cause of morbidity after aSAH, arising in up to one-third of patients presenting with aSAH.[30] Reversible constriction of the cerebral vasculature – vasospasm – occurs in two-thirds of patients presenting with aSAH.[31] While it is a recognised cause of DCI, recent effective treatments of vasospasm have not improved neurological outcomes.[32] The consequent interpretation is that ischaemia arises from other causes such as spreading cortical depolarisation or cerebral microcirculatory thrombosis, for which vasospasm may be just an epiphenomenon.[32] [33]

This does not mean that vasospasm should be ignored, but rather that its treatment should be focused on and titrated towards measures of functional neurologic outcome, rather than the surrogate features of vasospasm.[34]

Cerebral vasospasm occurs between days 3 and 14, following aSAH with a peak incidence at 7–10 days, with highest risks seen in those with large amounts of blood in the subarachnoid and intraventricular spaces.[35] Clinical presentation includes decreased level of consciousness, confusion and new focal neurologic deficit.

Consequently, clinical examination still offers the best combination of sensitivity and specificity to neurological deterioration.

Transcranial Doppler monitoring uses ultrasound assessment of cerebral arterial flow velocity to assess cerebral perfusion, with increasing confidence of vasospasm as velocity rises. The Lindegaard ratio is a comparison of velocities between the extracranial internal cerebral artery and ipsilateral middle cerebral artery. Ratios above six indicate severe spasm, while moderate vasospasm is associated with ratios over 3.[36]

Other modalities such as cerebral microdialysis and continuous electroencephalogram are employed but remain unproven modifiers of outcome.

With a premise that cerebral vasospasm reduces perfusion to the extent that it induces ischaemic change, treatment is based on the restoration of adequate levels of blood flow to meet demand. ‘Triple-H therapy’ (induced hypertension, induced hypervolaemia and haemodilution) has been practiced for over 40 years, but with emerging concern of complications and questionable efficacy.[37] [38] [39] Increasing practice has been refined to concentrate on induced hypertension in circumstances of euvolaemia. While this has been demonstrated to improve individual outcomes, effective pharmacological treatment of vasospasm has not improved neurologic outcomes, suggesting other additional mechanisms of injury.[32] [40]

Nimodipine has a modest reduction of the risk of poor outcome after aSAH but does not affect vasospasm.[41] [42] It may have fibrinolytic effects, reducing microcirculatory thrombosis, which is a postulated mechanism of cerebral ischaemia.[43] However, endovascular rescue therapy may be employed to treat cerebral vasospasm, delivering intra-arterial vasodilators by selective catheterisation, which may, in turn, allow balloon angioplasty of those vasospastic segments.[44] [45] [46] [47]



Neurological outcome is considerably modified by secondary insults involving other than the central nervous system, many of which offer potential for treatment or modification within the ICU.[48]

The ‘brain–heart axis’ is now well appreciated with acute neurologic disease processes manifesting changes upon cardiac function. These range from frequently observed changes in electrocardiogram (ECG) morphology and rhythm,[49] through abnormal perfusion scans for over 30% of SAH patients,[50] troponinemia in a similar proportion,[51] to abnormal echocardiography (8%–10% of SAH patients), most marked in those with basal cistern haemorrhage.[52] ECG changes have little effect on outcome, whereas troponinemia is associated with a higher incidence of reduction in ejection fraction and the presence of regional wall abnormalities.

This entity of neurologic ‘myocardial stunning’ is functionally related to takotsubo cardiomyopathy,[53] and similar to that entity, there is no indication for acute treatment of coronary ischaemia. Function recovers well within a matter of days. Both SAH and coronary artery disease share many risk factors including age, and the possibility of coincident acute coronary occlusion should not be totally dismissed.

The SAH patient typically presents with the worst headache of life – the pain and agitation of which usually indicate some intervention to limit blood pressure. The presence of pain and no requirement for blood pressure control should raise some suspicions of neurologic stunning – especially when accompanied by QTc prolongation, symmetrical T inversion and troponin elevation. Early echocardiography is indicated as this may determine whether surgical clipping or endovascular treatment is appropriate.[52]

Even after aneurysmal occlusion, myocardial stunning may induce a choice to proceed to early angioplasty as opposed to pressure elevation.[54]

Lung injury of some form is a frequent accompaniment in over 40% of patients with SAH, with around 12% experiencing severe effects.[55] Myocardial dysfunction may be associated, but there is potential induction of systemic inflammatory response with cerebral release of cytokines into the cardiopulmonary circulation.[56] This risk of lung injury is increased by transfusion.[57]

It is important to note that while worse neurological outcomes are seen in anaemic patients, transfusion is not associated with better outcomes but may rather worsen the risks.[58] Given that, many of the cardinal studies in ICU transfusion excluded neurological patients. There is no good body of evidence to drive practice – many (including my own unit) effect compromise with a transfusion haematocrit trigger of 25 as opposed to 21, employed in general ICU practice.

A significantly more prevalent risk factor is fever, with a discernible effect on outcome.[48] [59] [60] Fever correction to normothermia is the mainstay of treatment. Hypothermia has not shown benefit and complicates care.[61]

Fluid and electrolyte abnormalities are common in patients with SAH, with around 40% of patients experiencing hyponatraemia.[62] Argument continues as to whether the cause is cerebral salt wasting syndrome (CSWS)[63] [64] or the syndrome of inappropriate antidiuretic hormone.[65] This author has seen patients who demonstrate convincing evidence of both – but this is rare. Treatment differs according to cause with implications to volume status. The observed frequency of polyuria supports a higher incidence of CSWS, but this may be complicated by a natriuretic response to pressure elevation and volume loading.[66]

Enterally administered salt tablets are often enough to limit derangements – especially when combined with free water restriction. If there is a large urinary sodium loss, then fludrocortisone is indicated. Plasma and urine analysis will help decision-making. Hypertonic saline infusions may be indicated in circumstances of significant hyponatraemia below 130 mEq/L.

Many of these patients are anorexic, consequent to pain, nausea and opiate analgesia. There is a consequent risk of negative catabolism in the context of the significantly elevated energy requirements associated with SAH[67] – which in turn induces a negative nitrogen balance. That negative balance is associated with both hypoalbuminaemia – further reducing intravascular volume[68] – and increased risks of hospital-acquired infection and adverse outcome.[69]



SAH is an acute threat to life which, even after being treated, presents subsequent features of neurologic, cardiac, pulmonary, endocrine and nutritional dysfunction. This indicates a need for a comprehensive holistic appraisal of pathophysiology, which adds value to a dedicated neurocritical care team, a concept that has been validated by reductions of in-hospital mortality and length of stay.[70] [71] [72]

Financial support and sponsorship



Conflicts of interest

There are no conflicts of interest.


No conflict of interest has been declared by the author(s).

Address for correspondence:

Prof. Michael J. Souter
Department of Anesthesiology and Pain Medicine
University of Washington School of Medicine, Harborview Medical Center, 325 Ninth Avenue, Seattle, WA